flash memory cards also are somewhat limited in terms of data retention. So for long-term archieval purposes you'd still need to copy the files onto other storage media. For that purpose I'd suggest at least one USB hard drive plus M-Disc BD-Rs or two USB hard drives (ideally with two different HDD models).
And regarding the USB hard drives, mind you, that in the meantime there are two variants - one uses conventional SATA HDDs plus an additional interface with USB<->SATA bridge, the other HDDs with USB port. Personally I'd rather go for the former, because in case of USB port failure (e.g. due to ESD) one can still put the SATA HDD into another external drive case (or use it with one of those USB<->SATA dongles/docks...).
It is an interesting consideration, which can seem off-topicat first sight but, since we are talking about data preservation, a discussion is certainly worthwhile.
As a standard procedure, yes, it is a good idea to keep the original media where they belong, namely in the recorder. They have properties that make them particularly suitable for field recording (small size, convenience, low current draw, etc.) but they are not usually good candidates for long-term storage, if only because of the cost per megabyte and their comparatively low capacity per media, which would make library management impractical..
As a side note, it is a very bad idea to toss the originals after digitizing or copying them to the new fashionable format. The history of preservation is full of horror stories of entire collections lost because the copy had not been processed properly or had been recorded on media that proved defective only after the originals had been sent to the tip. I don't know if that was SalesOp's intention, but if it was, forget it!
Another all-important rule of data preservation is nicknamed 3-2-1: every file must exist in at least 3 copies, at least on 2 separate media, of which at least one is stored in a separate place so that it survives a fire (does Universal or OVH ring a bell?). This rule is the best protection we have to avoid problems, and it can be applied to every case, ranging from nationwide data repositories to home offices or your family archives. No need to store data on sophisticated RAID clusters (which have their usefulness too, just not for long-term archiving), you can save to lowly USB drives and be safe as long as they don't fail all at the same time - a very unlikely event with the 3-2-1 routine.
Of course it is also important to follow the technological evolution in order to save the files to the newer formats/interface standards, so that they are still playable when USB or the current file formats become obsolete. This takes more work than the 3-2-1, but is necessary if you plan to use your data in let's say, 20 years or more. The durability of a given media, such as self-erasure or chemical decay, is only secondary in this context, and having 3 copies in a format that you can no longer read would obviously be useless. Long-term storage is of primary importance for organizations that need durability, such as registers of birth for instance, and after a sizable number of catastrophic failures in the past they actually pay people to deal with all the associated issues.
On a smaller scale, if you use your data on a regular basis you often find yourself doing it almost inconsciously, as you quickly discover that open formats are better than proprietary formats when you have to move to new software, and as you copy your old files from obsolete parallel ATA or SCSI disks to SATA or USB disks for compatibility with your current computer and in the foreseeable future. It becomes instinctive.
So digitizing a cassette collection is a good idea, but it can be just a part of a an overall process that far exceeds the preservation of music. Of course you don't have to do the whole package, but it's a good opportunity to think about it.